HI,
A referee commented that the low number of trials in my study may render estimates of some beta weights unstable. On average, there are more than 18 trials per condition. Are these too few? How could I show that my signal is robust enough? Should I try and run the analyses excluding subjects with too few trials per condition (no of trials < 10? < 15?) and show that I yield comparable results?
Thanks for any advice.
Elisa